The Dells of Open Source

One common way to innovate is to combine two or more different things and create something new that is more than a sum of parts. This happens in music all the time. Musicians take what they like -- their influences -- and combine them in their own way. Sometimes you get literal combinations that create little genres like "trip-hop," "ambient dub," our "sadcore," but most often you just get bands that, "sound like a faster mix of band X and band Y."

Sooo... what do you get when you combine Dell and Open Source? As covered by CNET in Industry veterans bet on open-source model, you get a startup called SourceLabs. SourceLabs looks to make money by efficiently sourcing, pre-integrating, and selling solutions made from low-cost ($0) open source software packages. I like the analogy, and certainly being the "Dell" of anything has a nice ring to it.

Now this is nothing that Redhat and others haven't done for the likes of Linux and other now established open source products. So in a way, it's just more of the same good stuff -- giving corporate adopters of open source "one throat to choke." But still, I like thinking about this in terms of Dell, rather than just purely in terms of selling support and peace of mind. Dell makes money combining commodity hardware and offering computing solutions, Redhat does the same for operating system, and SourceLabs will pre-integrate and sell open source "stacks." No word yet on which or how many stacks.

My big question is how does this scale? Can SourceLabs, or someone like them, go big and become the Amazon of open source? Or just focus on a core set of applications, like Linux (or Dell)? I expect to see many more companies like this in the near future.

Open Source and the Transistor Radio?

I recently listened to Clayton Christensen's presentation from the 2004 Open Source Business Conference called "Capturing the upside." (Actually I streamed it from Doug Kaye's wonderful IT Conversations site).

Clayton compared the adoption and success of open source to Sony and their first transistor radios and the first Japanese autos exported to the US. Basically the early transistor radios were feature-wise pitiful when compared to the existing products on the market. However, they succeeded because they addressed a new market: young people who would put up with the static and poor reception because they could actually afford them, combined with the fact that they relished the ability to listen privately to there rock music, away from parental earshot. Existing radios, while high-quality, were big and expensive, and thus out of reach of most teenagers. So while the existing market supposedly clamored for ever-more features, the "good enough" transistor product proved that features and functionality had gotten ahead themselves, and the new "poor quality" upstart came in to open a open a new market. And then of course the transistor radio slowly improved, and the rest is history...

Clayton points out that open source is doing the same thing today. Beyond the now established open source infrastructure standouts like Linux, MySQL, Apache, etc., there's a newer crop of applications like SugarCRM, Gluecode, and Firefox that have also started out humbly as "good enough." Many people just aren't served by feature-saturated, high-priced commercial offerings. Instead they're finding relatively immature open source alternatives to be good enough, and also lower TCO or be more customizable or... I'm now looking at innovation and market evolution with respect to open source with a new twist -- thanks Clayton!

Angel funding coming back?

I went to SDForum's VC SIG tonight to hear an angel investor panel discussion on the climate for angel investment. The general consensus was that the angel community is slowly but surely moving out of the "fear" mode it has been in since the bubble burst and earlier angel rounds were getting "crammed down" when VC's came in on down rounds. Interestingly, the angel deals that have been getting done during the downturn have consisted of larger groups investing smaller amounts ($5-25k each).

But the bottom line is deal flow is supposedly picking up, and investments are getting done, mostly in the $1M-$3M pre-money valuation for prototype to nascent revenue companies (the panel admitted it's hard to generalize valuation). And the key to determining whether to fund the company: whether a VC will follow after new milestones are hit. The event was standing-room only and well run -- I look forward to going back again next month.

Minor Confusion over Open Source BPEL

Paul Brown sent me a heads up about the recent Infoworld article on open source BPEL (ObjectWeb plans open source BPEL server). The article also mentioned the recently announced Open Integration Suite, and mentioned that Orbeon had also recently contributed a BPEL Server to open source:

"It won't be the only open source BPEL server available. Orbeon Inc., a systems integrator in Mountain View, California, released the source code for its BPEL server a few weeks ago, along with several other components that comprise the Orbeon Integration Suite."

Unfortunately that's not quite the case... Orbeon did found and seed the Open Integration Suite project with a contribution of it's OXF product line, but it will be sourcing other complimentary components, including a BPEL Server, from other open source projects. The good news is that eMaxx's is now a candidate for inclusion in the suite! (Disclaimer: I am currently on Orbeon's board of advisors.)